---small_w_tester.nlogo

NETLOGO

to move_policies
ifelse ((distance turtle p-target) > (distance turtle p-source))
[fd (0.5 * ( 1 / link_length) * (distance turtle p-target)) + 1]
[fd (distance turtle p-target)]
end

to doPlots
if clustcoeff? [set-current-plot "Nodes"
    set-current-plot-pen "clustcoeff"
    set-plot-pen-mode 2
        plotxy p_rand_log ( ((last cclist) - (last cc_p1_list)) / ((last cc_min_list) - (last cc_p1_list)) )]
if pathlen?   [     set-current-plot "Nodes"
    set-current-plot-pen "pathlen"
      set-plot-pen-mode 2
        plotxy p_rand_log ( ((last pathllist) - (last pathl_p1_list)) / ((last pathl_min_list) - (last pathl_p1_list)) )]
if pol_action?  [  set-current-plot "Nodes"
    set-current-plot-pen "polsteps"
    set-plot-pen-mode 2
        plotxy p_rand_log ( ((last polstepslist) - (last polsteps_p1_list)) / ((last polsteps_min_list) - (last polsteps_p1_list)) )]
end


to dopolplots
set-current-plot "poldev"
   
            plotxy polsteps 100 * (count actors with [color = red]) / num_actors
end
@#$#@#$#@
GRAPHICS-WINDOW
490
10
801
342
150
150
1.0
1
10
0
1

CC-WINDOW
151
10
479
295
Command Center

BUTTON
4
10
85
43
NIL
setup
NIL
1
T
OBSERVER

BUTTON
4
45
85
78
NIL
step
NIL
1
T
OBSERVER

BUTTON
4
79
85
112
go
while [repcounter <= reps]\n[step]
NIL
1
T
OBSERVER

MONITOR
7
526
64
575
NIL
steps
3
1

SWITCH
595
350
719
383
drawLabels?
drawLabels?
1
1
-1000

PLOT
151
305
481
584
Nodes
log p
norm var
-3.0
0.2
0.0
1.0
true
true
PENS
"clustcoeff" 1.0 2 -8716033 true
"pathlen" 1.0 2 -65413 true
"polsteps" 1.0 2 -11352576 true
"Four" 1.0 0 -16711681 true
"Five" 1.0 0 -16711936 true
"Six" 1.0 0 -6524078 true
">Six" 1.0 0 -16777216 true

SLIDER
3
246
148
279
num_actors_start
num_actors_start
5
150
25
5
1
NIL

SLIDER
3
319
146
352
p_rand_log_start
p_rand_log_start
-4
0
-3.0
0.2
1
NIL

SWITCH
491
350
589
383
drawnet?
drawnet?
0
1
-1000

SLIDER
3
396
145
429
link_length
link_length
1
5
1
1
1
NIL

MONITOR
74
526
131
575
tol
tol
3
1

SLIDER
3
434
144
467
range
range
0
50
4
1
1
NIL

MONITOR
600
389
657
438
av l(p)
short_l
3
1

MONITOR
625
442
682
491
clustc
clust_coeff
3
1

SLIDER
3
281
147
314
num_actors_end
num_actors_end
5
150
35
5
1
NIL

SLIDER
3
357
143
390
p_rand_log_end
p_rand_log_end
-4
0
0.0
0.2
1
NIL

MONITOR
814
10
891
59
NIL
num_actors
3
1

MONITOR
815
67
865
116
p
10 ^ p_rand_log
4
1

SLIDER
4
472
143
505
start_policy_%
start_policy_%
0
100
60
2
1
NIL

PLOT
690
393
990
586
poldev
iter
%_pol_agree
0.0
20.0
0.0
100.0
true
false

MONITOR
488
533
565
582
red actors
count actors with [color = red]
0
1

MONITOR
579
533
648
582
NIL
polsteps
0
1

SWITCH
490
388
596
421
pathlen?
pathlen?
0
1
-1000

SWITCH
490
442
613
475
clustcoeff?
clustcoeff?
0
1
-1000

SWITCH
489
497
617
530
pol_action?
pol_action?
0
1
-1000

MONITOR
905
235
968
284
NIL
shortcuts
3
1

MONITOR
815
123
950
172
NIL
100 * (10 ^ p_rand_log)
3
1

MONITOR
815
179
874
228
NIL
randval
3
1

MONITOR
899
10
949
59
edges
num_actors * range
3
1

MONITOR
815
235
900
284
exp. shortcuts
(10 ^ p_rand_log) * num_actors * range 
3
1

MONITOR
815
291
929
340
NIL
length edge_list
3
1

MONITOR
877
67
950
116
NIL
p_rand_log
3
1

SLIDER
5
145
147
178
reps
reps
0
100
3
1
1
NIL

MONITOR
36
182
108
231
repetition
repcounter
0
1

@#$#@#$#@
Small world parameters and political activity in a network

Network regular and randomized according to rules in Watts and Strogatz 1998

somewhere in between the regular and randomised networks we should see a small world, if parameters are right, this is recognised by the large drop in global path length withpout a corresponding fall in clustering.

We can examine alongside this an actor based simulation of political activity in networks. At present this is a (not very good) rule that looks towards convergence of opinion on a policy, where actors are influenced by their inetractions with each other. I have a paper in progresson this subject. I can make drafts available if people are interested.

What we hope to see is that some kind of believable model of policy formation in a network is affected by the small world properties eing present.

Things to note
1) depending on the size of your network/ expected number of shortcuts you can boost up the minimum p-log value and save a load of worthless iterations
2) the poldev graph is not much use (at the momnet), it just shows that the policy simulation is alive and kicking
3